Anderson et al.: Evolutionary associations between Cynoscion arenarius and C. nothus 
17 
Table 2 
Morphometric and meristic differences between two seatrout species, sand seatrout ( Cynoscion arenarius) and silver seatrout (C. 
nothus), collected in trawls from Galveston Bay, TX, in July 2007. The range includes the overall range of data observations (n), 
followed by the mean and SE ( standard error). The significance of each comparison was calculated by using either a standardized 
t-test (indictated as a lower-case a) or a chi-square test of homogeneity (indicated as a lower-case b). 
C. arenarius C. nothus 
n 
Range 
Mean ±SE 
n 
Range 
Mean SE 
P-value 
Size measurements 
Standard length (mm) 
60 
83-175 
107.9 ±8.3 
60 
100-163 
119.5 ±4.7 
<0.0001 a 
Body weight (g) 
60 
10-95 
26.4 ±7.3 
60 
20-75 
33.4 ±4.1 
0.0015 a 
Ratios 
Pectoral fin/pelvic fin 
60 
0.9-1. 2 
1.10 ±0.04 
48 
0.9-1. 3 
1.09 ± 0.04 
0.5978 a 
Anal-fin base/eye diameter 
60 
1. 2-2.4 
1.59 ± 0.10 
60 
0.9-1. 2 
1.01 ± 0.03 
<0.0001 a 
Meristics 
Anal-fin soft rays 
60 
10-12 
10.9 ± 0.2 
60 
8-9 
8.8 ± 0.2 
<0.0001 b 
Lateral-line scales 
57 
57-61 
59.3 ± 0.6 
56 
57-63 
59.5 ± 0.8 
0.9151 b 
frequencies were assumed to be independent between 
populations. These model parameters were tested with 
a set K = 2, representing the two species as possible 
genetic contributors for each individual. The program 
was run under these model conditions for six trials to 
check for stability of resulting admixture coefficients 
(Q-values). 
In order to evaluate the significance of individuals 
with extreme admixture coefficients values of Q , a set 
of simulated populations of 1000 individuals from each 
species was generated from allele-frequency data with 
Whichloci. We analyzed these populations with Struc- 
ture using identical model parameters to those in the 
experimental populations; the probability of obtain- 
ing a higher estimated level of admixture for any test 
individual was estimated as the frequency of higher 
admixture scores in the simulated population. 
Results 
Sample statistics and morphological characters 
Samples of sand seatrout were collected from a com- 
bination of three grids, each of which was roughly 
one km from shore. The three grids were located at 
depths of three (one grid) and four (two grids) fathoms. 
For comparative purposes, trawl data for each species 
was combined and treated as a single random sample. 
All individual silver seatrout were collected from a 
single grid offshore from Galveston Bay. The grid was 
two km from land and had a depth of seven fathoms. 
The sand seatrout sample contained specimens that 
were significantly smaller than those obtained in the 
silver seatrout sample, in both mean standard length 
(t=— 4.74, P<0.0001) and mean weight (t=- 3.27, P=0.0015) 
(Table 2). The difference in mean size was not likely 
caused by gear selectivity because the size range of both 
species combined was from 83 to 175 mm, and larger fish 
are routinely caught in trawls. The mean ratio of weight 
to length was not significantly different between the spe- 
cies (t= 0.07, P=0.800), indicating a similarity in growth 
trajectories between the species at the size examined, 
despite significant differences in overall size. 
Two of four meristic measurements were useful in 
reliably sorting specimens to species. First, sand seat- 
rout had overall larger anal-fin to eye-diameter ratios 
(f=21.32, P<0.0001); the sand seatrout ratio ranged from 
1.23 to 2.44, whereas the range in silver seatrout was 
0.85-1.19. Second, anal-fin soft rays were significantly 
different between species (x 2 = 120 , df=10, P<0.0001); 
sand seatrout possessed an average of 10.9, silver seat- 
rout an average of 8.8 soft rays. In contrast, there was 
not a significant difference in pectoral-fin to pelvic-fin 
ratios between species (t= 0.53, P=0.5978) (Fig. 3), nor 
was there a significant difference in the number of lat- 
eral line scales (^=7.47, df=14, P=0.915). The anal-fin 
soft-ray meristic was the most practical morphological 
character for species discernment because the range 
of this character did not overlap between species in 
any specimen (range 10-12 for sand seatrout, 8-9 for 
silver seatrout), and this character was relatively easy 
to count. 
Mitochondrial DNA 
The fragments recovered from each mtDNA amplifica- 
tion were approximately 1500 bp, and this size did not 
vary between species. Two distinctive RFLP patterns 
were identified (Fig. 4). The first pattern contained four 
bands, at approximately 450, 290, 250, and 190 bp. This 
pattern was identified in each of the 60 sand seatrout 
assayed. The second pattern also contained four bands, 
at approximately 400, 290, 200, and 190 bp. This pattern 
was identified in each of the 60 silver seatrout assayed. 
Based upon the expected relative intensity of each band. 
